We need a new generation of traffic modeling tools.
Looking at the cable downstream spectrum these days is like looking at an all-you-can-eat buffet of services. There’s something for everybody: A small number of remaining analog channels, SD channels, HD channels, 3-D channels, video-on-demand channels, switched digital video channels, high-speed data, VoIP and, in some cases, even legacy circuit-switched telephony channels.
But as much as cable already offers, more is on the horizon. While the exponential growth in bandwidth needed for HSD has been relatively steady, the drive for more business customers and video on IP devices such as computers, smartphones and tablet PCs is ratcheting up the requirements for HSD bandwidth faster than Nielsen’s law would have predicted. Interactive TV made possible by EBIF will put more pressure on downstream bandwidth requirements. And given the push of TV manufacturers to see new 3-D and connected TVs over the past year or two, we’ll most likely see a push toward 2k and even 4K ultra-high-definition (UHD) formats in the decade to come.
So how do the cable operators keep up with these increasing demands on downstream spectrum? The answer is a full toolbox of methods at the disposal of the cable operator. Analog reclamation can be done either gradually as more homes switch to digital TVs and/or by putting set-top boxes on every TV, or more quickly by deploying digital terminal adapters (DTAs), which are low-cost QAM receivers designed to convert SD versions of the former analog lineup back into analog signals for legacy NTSC TVs.
But there are many more tools at the operators’ disposal: SDV allows long-tail content to be sent over the downstream at the subscriber’s request. Modern HD STBs also support MPEG-4 compression, which can reduce the bandwidth required to send HD channels from an average of 15 Mbps (in an MPEG-2 format) to less than 8 Mbps for similar video quality. DOCSIS 3.0 provides statistical multiplexing gains from the combination of multiple HSD channels into a single, fat QAM channel. And variable and adaptive bit rate video transmission means better statistical multiplexing and grooming of channel multiplexes.
Each one of these techniques by themselves can improve the efficiency of downstream spectrum requirements considerably, but together they give cable operators the ability to optimize their spectral utilization over wide ranges of content needs, and even dynamically if needed. For example, QAMs might be used to serve business customers during the daytime and then be dual-purposed to serve residential VOD and SDV channels in the evening. More generally, the next-generation access architecture (NGAA) and converged multiservice access platform (CMAP) will create even greater QAM densities and flexibility in the network to seamlessly swap channel allocations between SD, HD, 3-D, VOD, SDV, HSD and the UHD that is on the horizon. Cable has plenty of tools in its downstream toolbox for adapting to new content types, mixes and dynamics.
But with all of these tools and tactics for adapting to downstream needs, it is becoming even more important to develop better traffic management models for the entire network. The Erlang theory was fine for telephone networks that had a single type of service and signaling on them, but modern cable downstreams will need a new generation of traffic modeling and forecasting tools.
What’s needed are solutions that integrate the existing models for individual services into a unified view of the entire cable network, from the national headend through the backbone and into the individual markets and access networks. Even the home network needs to be taken into consideration, especially if it is managed by the cable operator.
Here’s an example of how things have changed: Mother’s Day used to be the peak usage point for telephone networks; now it’s call-in TV programs such as “American Idol” and “Dancing With the Stars.” And Internet traffic peaks and blocking are even less predictable because of such factors as breaking news, the so-called Slashdot effect (where a larger website links to a smaller site that can’t handle the traffic), and malware and/or denial of service (DoS) or distributed denial of service (DDoS) attacks.
So as cable enters the all-digital era, the work continues to understand the variability and adaptability of our infrastructures. At SCTE, for example, we’ve created an extensive library of traditional courses, Webinars and primers that are focused on the challenges and potential of all-digital networks. In addition, we’ve created a new online course in association with Trilithic on the “Impact of going all-digital.”
Going all-digital is just the first step. The ultimate goal is to support operators and their engineering and operations teams to get the most out of their new all-digital architectures. By building a knowledge base now, engineering and operations teams will be fully prepared to deliver the new services that can drive subscriber growth and retention and build revenue for operators.